In a distributed processing environment, computational tasks may be assigned to a plurality of nodes or processes for execution. A central server may manage the scheduling of the system such that the processing ability of each node that executes a task may be fully exploited and the load of computations of the node may be optimized so that each task may be executed efficiently. However, improvements are needed to address inefficiencies which arise when processes abruptly stop running (i.e., go “down”) or when “up”/“down” messages broadcast from processes are not received by other processes.